Aktu Btech Machine Learning Techniques KCS-055 Short Question Notes Pdf

Learn more about the Machine Learning Techniques Short Question Notes from the B.Tech. AKTU Quantum Book. Discover techniques and algorithms for intelligent data analysis, forecasting, and decision-making across a range of fields.

Dudes 🤔.. You want more useful details regarding this subject. Please keep in mind this as well.

Important Questions For Machine Learning Techniques:
*Quantum               *B.tech-Syllabus
*Circulars                *B.tech AKTU RESULT
* Btech 3rd Year     * Aktu Solved Question Paper

Unit-I: Introduction

Q1. Define machine learning. 

Ans. Machine learning is an artificial intelligence application that allows computers to automatically learn and improve from experience without being explicitly designed. 

Q2. What are the different types of machine learning algorithm? 

Ans. Different types of machine learning algorithm are:

• 1. Supervised machine learning algorithm 
• 2. Unsupervised machine learning algorithm 
• 3. Semi-supervised machine 
• 4. learning algorithm Reinforcement machine learning algorithm.

Q3. What are the applications of machine learning ? 

Ans.  Applications of machine learning are : 

• 1. Image recognition 
• 2. Speech recognition 
• 3. Medical diagnosis 
• 4. Statistical arbitrage 
• 5. Learning association  

Q4. What are the advantages of machine learning?

Ans. Advantages of machine learning: 

• 1. Quickly discovers trends; and
• 2. No patterns require human intervention.
• 3. Constant improvement.
• 4. Dealing with multidimensional and multivariate data.

Q5. What are the disadvantages of machine learning ?

Ans. Disadvantages of machine learning:  

• 1. Data acquisition
• 2. Time and resources 
• 3. Interpretation of results 
• 4. High error-susceptibility 

Q6. What is the role of machine learning in human life ?

Ans. Role of machine learning in human life: 

• 1. Learning 
• 2. Reasoning 
• 3. Problem solving 
• 4. Language understanding 

Q7. What are the components of machine learning system ? 

Ans. Components of machine learning system are :

• 1. Sensing 
• 2. Segmentation 
• 3. Feature extraction 
• 4. Classification 
• 5. Post processing  

Q8. What are the classes of problem in machine learning?  

Ans. Classes of problem in machine learning are:

• 1. Classification 
• 2. Regression 
• 3. Clustering 
• 4. Rule extraction 

Q9.  What are the issues related with machine learning? 

Ans. Issues related with machine learning are: 

• 1. Data quality 
• 2. Transparency 
• 3. Traceability 
• 4. Reproduction of results

Q10. Define supervised learning. 

Ans. Supervised learning, also known as associative learning, involves training the network by feeding it input and matching output patterns.

Q11. Define unsupervised learning ?

Ans. Unsupervised learning, often known as self-organization, is the process by which an output unit is trained to respond to clusters of patterns in the input.  

Q12.  Define well defined learning problem. 

Ans. When a computer programme learns from experience E in relation to a class of tasks and performance measure P, its performance at tasks in T, as measured by P, improves with experience E.  

Q13. What are the features of learning problems ?  

Ans. Features of learning problems are : 

• 1. The class of tasks (T). 
• 2. The measure of performance to be improved (P). 
• 3. The source of experience (E).

Q14.  Define decision tree learning.  

Ans. Predictive modeling methodologies such as decision tree learning are utilized in statistics, data mining, and machine learning. It use a decision tree to progress from observations about an object to conclusions about the target values of the item.

Q15. What is decision tree ?  

Ans. A decision tree is a decision-making tool that employs a tree-like model of decisions and their potential outcomes, such as chance event outcomes, resource costs, and utility. 

Q16. What are the types of decision tree? 

Ans. There are two types of decision tree: 

• 1. Classification tree 
• 2. Regression tree  

Q17. Define classification tree and regression tree.

Ans. Classification tree: A classification tree is an algorithm with a fixed target variable. This algorithm is used to determine which class a target variable belongs to. 

Regression tree: A regression tree is an algorithm that predicts the value of a variable whose value is not fixed.

Q18. Name different decision tree algorithm. 

Ans. Different decision tree algorithms are: 

• 1. ID3 
• 2. C4.5 
• 3. CART

Q19. What are the issues related with the decision tree ? 

Ans. Issues related with decision tree are: 

• 1. Missing data 
• 2. Multi-valued attribute 
• 3. Continuous and integer valued input attributes 
• 4. Continuous-valued output attributes 

Q20. What are the attribute selection measures used in decision tree ? 

Ans. Attribute selection measures used in decision tree are:  

• 1. Entropy 
• 2. Information gain 
• 3. Gain ratio 

Unit-II: Regresion

Q1. Define the term regression. 

Ans. Regression is a statistical tool used in finance, investment, and other fields to identify the degree and character of a relationship between one dependent variable and a set of other factors (known as independent variables). 

Q2. What are the types of regression ?

Ans. Following are the types of regression :

• 1. Linear regression 
• 2. Logistic regression 

Q3. Define logistic regression.  

Ans. Logistic regression is a classification algorithm that uses supervised learning to predict the likelihood of a target variable. Because the nature of the target or dependant variable is dichotomous, there are only two viable classes.

Q4. What are the types of logistic regression ? 

Ans. Following are the types of logistic regression: 

• 1. Binary or Binomial logistic regression 
• 2. Multinomial logistic regression 
• 3. Ordinal logistic regression 

Q5. Define Bayesian decision theory. 

Ans. Bayesian decision theory is a fundamental statistical method to pattern recognition. This method is centred on quantifying the tradeoffs between various classification judgements by utilizing the likelihood and costs associated with such decisions. 

Q6. Define Bayesian belief network.

Ans. Joint conditional probability distributions are specified using Bayesian belief networks. Belief Networks, Bayesian Networks, and Probabilistic Networks are other names for them.

Q7. Define EM algorithm. 

Ans. The Expectation-Maximization (EM) algorithm is an iterative method for determining maximum-likelihood estimates for model parameters when the data is incomplete, has missing data points, or contains hidden variables. 

Q8. What are the usages of EM algorithm ?

Ans. Usages of EM algorithm are:

• 1. It can be used to fill in gaps in a sample’s data.
• 2. It can be utilized as the foundation for unsupervised cluster learning.
• 3. It can be used to calculate the parameters of the Hidden Markov Model (HMM).
• 4. It can be used to find latent variable values.

Q9. What are the advantages of EM algorithm ?

Ans. Advantages of EM algorithm are: 

• 1. Each cycle guarantees that the likelihood will increase.
• 2. The E-step and M-step are simple to apply.
• 3. There are closed-form solutions to the M-steps.

Q10. What are the disadvantages of EM algorithm ?

Ans. Disadvantages of EM algorithm are:

• 1. Its convergence is sluggish.
• 2. It only converges to the local optima.
• 3. It necessitates both forward and backward probabilities (numerical optimisation requires only forward probability).

Q11. Define support vector machine.

Ans. A support vector machine is a supervised machine learning method that examines data and sorts, analyses, and classifies it. 

Q12. What are the types of support vector machine ?  

Ans. Types of support vector machine are: 

• 1. Linear support vector machine 
• 2. Non-linear support vector machine 

Q13. What are the applications of SVM ?

Ans. Applications of SVM: 

• 1. Text and hypertext classification 
• 2. Image classification 
• 3. Recognizing handwritten characters 
• 4. Biological sciences, including protein classification

Unit-III: Decision Tree Learning

Q1. What is instance-based learning? 

Ans. Instance-Based Learning (IBL) is a classification algorithm that is an extension of nearest neighbour or KNN classification algorithms that does not keep a collection of abstractions of models produced from examples.

Q2. What are the advantages of KNN algorithm ?

Ans. Advantages of KNN algorithm are:

• 1. Notraining period. 
• 2. Because the KNN algorithm does not require training before making predictions, new data can be supplied without affecting the system’s accuracy. 
• 3. KNN is easy to implement. 

Q3. What are the disadvantages of KNN algorithm ?  

Ans. Disadvantages of KNN algorithm are:

• 1. It is ineffective with large datasets.
• 2. It is ineffective with large size.
• 3. It is necessary to scale the features.
• 4. It is susceptible to noise, missing values, and outliers.

Q4. Define locally weighted regression.

Ans. Locally Weighted Regression (LWR) is a memory-based approach that performs regression around a point of interest using local training data. 

Q5. Define radial basis function.

Ans. The value produced by a Radial Basis Function (RBF) is always an absolute value, meaning that it is a measure of distance and cannot be negative. It is a function that assigns a real value to each input from its domain (it is a real-value function). 

Q6. Define case-based learning algorithms. 

Ans. Case-based learning algorithms use a series of training instances as input and provide a concept description that can be used to predict the values of objective feature for cases that will be presented later. 

Q7. What are the disadvantages of CBL (Case-Based Learning) ?

Ans. Disadvantage of case-based learning algorithm: 

• 1. They use a lot of computing power to run because they save and analyse data from all training cases.
• 2. They can’t stand noise and pointless details.
• 3. They depend on the choice of the similarity function used by the algorithm.
• 4. They cannot process symbolic feature values in an easy way.

Q8. What are the functions of CBL ?

Ans. Functions of case-based learning algorithm are:

• 1. Pre-processor 
• 2. Similarity 
• 3. Prediction 
• 4. Memory updating  

Q9. What are the processing stages of CBL ? 

Ans. Case-based learning algorithm processing stages are: 

• 1. Case retrieval 
• 2. Case adaptation 
• 3. Solution evaluation 
• 4. Case-base updating

Q10. What are the benefits of CBL as lazy problem solving method ? 

Ans. The benefits of CBL as a lazy Problem solving method are: 

• 1. Ease of knowledge elicitation. 
• 2. Absence of problem-solving bias. 
• 3. Incremental learning. 
• 4. Suitability for complex and not-fully formalized solution spaces. 
• 5. Suitability for sequential problem solving. 
• 6. Ease of explanation. 
• 7. Ease of maintenance. 

Q11. What are the applications of CBL ? 

Ans. Applications of CBL: 

• 1. Interpretation 
• 2. Classification 
• 3. Design 
• 4. Planning 
• 5. Advising 

Q12. What are the advantages of instance-based learning ? 

Ans. Advantages of instance-based learning: 

• 1.  Learning is trivial 
• 2. Works efficiently 
• 3. Noise resistant 
• 4. Rich representation, arbitrary decision surfaces 
• 5. Easy to understand 

Q13. What are the disadvantages of instance-based learning ? 

Ans. Disadvantages of instance-based learning: 

• 1. Need lots of data. 
• 2. Computational cost is high. 
• 3. Restricted to x ∊ Rⁿ. 
• 4. Implicit weights of attributes (need normalization). 
• 5. Need large space for storage i.e., require large memory. 
• 6. Expensive application time.

Unit-IV: Artificial Neural Network

Q1. What are neurons ? 

Ans. A neuron is a tiny cell that receives different electrochemical signals from different sources and responds by sending electrical impulses to other neurons. 

Q2. What is artificial neural network ? 

Ans. Artificial neural networks are computational algorithms designed to mimic the behaviour of neuron-based biological systems.

Q3. Give the difference between supervised and unsupervised learning in artificial neural network. 

Ans. 

Q4. Define activation function. 

Ans. The foundational component of a brain model is an activation function. It is employed to control the amplitude of a neuron’s output. Also known as the squashing function. 

Q5. Give types of activation function. 

Ans. Types of activation function: 

• 1. Signum function
• 2. Sigmoidal function 
• 3. Identity function 
• 4. Binary step function 
• 5. Bipolar step function  

Q6. Give advantages of neural network. 

Ans. Advantages of neural network:

• 1. A neural network is capable of carrying out tasks that a linear programme is unable to.
• 2. It is adaptable to all applications.
• 3. A neural network picks up new information without having to be reprogrammed.

Q7. What are disadvantages of neural network (NN) ?
Ans. Disadvantages of neural network:

• 1. To function, the neural network needs to be trained.
• 2. It takes a long time to process a large NN.

Q8. List the various types of soft computing techniques and mention some application areas for neural network.  

Ans. Types of soft computing techniques: 

• 1. Fuzzy logic control 
• 2. Neural network 
• 3. Genetic algorithms 
• 4. Support vector machine

Application areas for neural network:  

• 1. Speech recognition 
• 2. Character recognition 
• 3. Signature verification application 
• 4. Human face recognition  

Q9. Draw a biological NN and explain the parts. 

Ans. 1. The peripheral nervous system connects actual biological neurons to form biological neural networks.

2. A biological neural network is often made up of a collection of chemically linked or functionally related neurons.

A biological neural network has three major parts:

• 1. Soma or cell body: The nucleus of the cell is inside, along with other important organelles that carry out specific functions. 
• 2. A set of dendrites: It grows out from the cell in the shape of a tree. Along these dendrites, the neuron gets its input electrical signal.
• 3. Axon: An electrical signal is transferred from the cell (Soma) to another neuron for processing via tabular extension.  

Q10. What is single layer feed forward network ?

Ans. The simplest type of layered network is a single layer feed forward network, in which an input layer of source nodes projects onto an output layer of neurons but not the other way around. 

Q11. Write different applications of neural networks (NN). 

Ans. Applications of NN are : 

• 1. Image recognition 
• 2. Data mining 
• 3. Machine translation 
• 4. Spell checking 
• 5. Stock and sport bet prediction 
• 6. Statistical modeling 

Q12. Draw an artificial neural network.

Ans. A computational model called an Artificial Neural Network (ANN) is based on the composition and operation of biological neural networks.

Q13. What do you mean by neural network architecture ? 

Ans. The organization of neurons into layers, as well as the connections between layers, activation processes, and learning strategies, are referred to as neural network architecture. How a network converts its input into an output depends on its design and neural network model.

Q14. What are the types of neuron connection ? 

Ans. Following are the types of neuron connection :

• 1. Single-layer feed forward network 
• 2. Multilayer feed forward network 
• 3. Single node with its own feedback 
• 4. Single-layer recurrent network 
• 5. Multilayer recurrent network

Q15. What is gradient descent ?

Ans. Iteratively traveling in the direction of the steepest descent, as indicated by the negative of the gradient, gradient descent is an optimisation process used to minimize some function.

Q16. What are the types of gradient descent ?  

Ans. Types of gradient descent are:

• 1. Batch gradient descent 
• 2. Stochastic gradient descent 
• 3. Mini-batch gradient descent 

Q17. What is self organizing map (SOM)? 

Ans. 1. Self Organizing Map (SOM) is a data visualization technique that reduces the dimensions of data to a map, making it easier to grasp high dimensional data.

2. SOM illustrates the clustering notion by combining data that are related.

Unit-V: Reinforcement Learning

Q1. Define genetic algorithm. 

Ans. The mechanics of natural genetics and natural selection are the foundation of genetic algorithms, which are computerised search and optimisation algorithms. These algorithms develop search and optimisation procedures by imitating the natural selection and genetics principles.  

Q2. Give the benefits of genetic algorithm.  

Ans. Benefits of genetic algorithm are:

• 1. They are Robust. 
• 2. They offer big space state optimisation.
• 3. They do not malfunction in the presence of noise or a little change in input.

Q3. What are the applications of genetic algorithm?

Ans. Following are the applications of genetic algorithms:  

• 1 Recurrent neural network 
• 2. Mutation testing 
• 3. Code breaking 
• 4. Filtering and signal processing 
• 5. Learning fuzzy rule base 

Q4. What are the disadvantages of genetic algorithm ?

Ans. Disadvantages of genetic algorithm: 

• 1. As the fitness function relies on the issue, identification is challenging.
• 2. It can be challenging to choose the best genetic operators.

Q5. Define genetic programming. 

Ans. Machine learning is a subset of evolutionary algorithms, which includes genetic programming (GP). EAs are employed to find solutions to issues that humans are unable to handle.

Q6. What are the advantages of genetic programming ? 

Ans. Advantages of genetic programming are:

• 1. In GP, the algorithm can create an enormous number of different programmes.
• 2. Even though GP employs machine code, which aids in producing results very quickly, using any high-level language that requires compilation can produce errors and slow down our programme.
• 3. There is a strong likelihood that even a very slight variation will have a devastating impact on the fitness of the created solution.

Q7. What are the disadvantages of genetic programming ?  

Ans. Disadvantages of genetic programming are: 

• 1. The maximum length can be increased up to hardware limitations because it does not impose any predetermined length of solutions.
• 2. In genetic programming, it is not required for a person to possess the greatest possible understanding of the issue and its remedies.

Q8. What are different types of genetic programming ? 

Ans. Different types of genetic programming are : 

• 1. Tree-based genetic programming 
• 2. Stack-based genetic programming 
• 3. Linear genetic programming 
• 4. Grammatical evolution 
• 5. Cartesian Genetic Programming (CGP) 
• 6. Genetic Improvement Programming (GIP) 

Q9. What are the functions of learning in evolution ? 

Ans. Function of learning in evolution :  

• 1. It enables people to adapt to environmental changes that happen over the course of a lifetime or a few generations.
• 2. It channels evolutionary search by allowing evolution to use data gathered from the environment.
• 3. It can assist with and direct evolution.

Q10. What are the disadvantages of learning in evolution ? 

Ans. Disadvantages of learning in evolution are :

• 1. A delay in the ability to acquire fitness. 
• 2. Increased unreliability. 

Q11. Define learnable evolution model.

Ans. A non-Darwinian approach to evolutionary computation called the Learnable Evolution Model (LEM) uses machine learning to direct the creation of new individuals (candidate problem solutions).

Q12. What are different phases of genetic algorithm ?

Ans. Different phases of genetic algorithm are:

• 1. Initial population 
• 2. FA(Factor Analysis) fitness function 
• 3. Selection Crossover 
• 5. Mutation 
• 6. Termination  

Q13. Define sequential covering algorithm. 

Ans. Generally speaking, a sequential covering algorithm learns a single rule repeatedly to produce a decision list (or set) that covers the entire dataset rule by rule.

Q14. Define Beam search.

Ans. A heuristic search strategy called beam search expands the most promising node within a small group to explore a graph.

Q15. What are the properties of heuristic search?

Ans. Properties of heuristic search are:

• 1. Admissibility condition 
• 2. Completeness condition 
• 3. Dominance properties 
• 4. Optimality property 

Q16. What are different types of reinforcement learning? 

Ans. Different types of reinforcement learning are:

• 1. Positive reinforcement learning 
• 2. Negative reinforcement learning  

Q17. Define Q-learning. 

Ans. Q-learning is model-free reinforcement learning, which is typically easier to implement. Reinforcement learning is the problem that an agent must solve when it needs to learn behaviour through trial-and-error interactions with a dynamic environment.  

Q18. Define positive and negative reinforcement learning. 

Ans. Positive reinforcement learning: Positive reinforcement learning is defined as when an event happens as a result of a certain behaviour, such as when the frequency and strength of the behaviour grow.

Negative reinforcement learning: Positive reinforcement refers to the strengthening of a behaviour as a result of stopping or avoiding a negative circumstance.

Machine Learning Techniques Btech Quantum PDF, Syllabus, Important Questions

Machine Learning Techniques Quantum PDF | AKTU Quantum PDF:

AKTU Important Links | Btech Syllabus

Important Links-Btech (AKTU)